The numbers are surprisingly low if accepted at face value, but I do not think that any of us believes that the actual number of " mistake kills" is the same as the reported number of mistake kills.